English subtitles

← Why Impute - Intro to Data Science

Get Embed Code
4 Languages

Showing Revision 5 created 05/24/2016 by Udacity Robot.

  1. In scenarios where we don't have very much data, or
  2. where removing our missing values would compromise the representativeness of
  3. our sample, it might not make sense to throw away
  4. a bunch of our entries just because they're missing values.
  5. This could severely impact the statistical power of whatever analysis
  6. we were trying to perform. In this case, it likely
  7. makes sense to make an intelligent guess at the missing
  8. values in our data. The process of approximating these missing
  9. values is referred to as imputation. There are many
  10. different ways to impute missing values. And different techniques
  11. are constantly being developed. I want to quickly discuss
  12. some relatively simple ways to impute missing values in our
  13. data. Let's note that imputation is a really hard
  14. problem. Each of the methods we'll discuss introduce a
  15. certain biases or inaccuracies into your data set. We're
  16. discussing some of the most simple ways to impute data,
  17. but much more sophisticated and robust methods are out there.